Impact of Globodera ellingtonae on yield of potato (Solanum tuberosum)

Field experiments

Field trials were carried out in 2013 (two trials), 2014 (two trials), and 2015 (one trial) at OSU-COARC, Powell Butte, OR (Table 1). The field used in all years and trials had been in a long-term four-year potato rotation. It was planted to barley (Hordeum vulgare) and Austrian winter pea (Pisum sativum) in 2007, to potato in 2008, and was fallow in 2009 and 2010. At the end of summer 2010, the field was planted to a winter oat (Avena sativa) cover crop and in all subsequent winters was cover cropped with winter wheat (Triticum aerstivum). Each year, trials were placed in an area of the field where pretreatment soil sampling did not recover G. ellingtonae cysts. The soil at Powell Butte, OR is a Redmond ashy sandy loam.

In each year, an area approximately 20-m wide × 45-m long was prepared for potato planting. Treatments of varying Pi (Table 1) were applied to individual potato plants using cysts reared as described above. In 2014, a trial was conducted to evaluate additional potato cultivars from different maturity classes two early (‘Russet Norkotah’ and ‘Yukon Gold’), one mid (‘Ranger Russet’), and two late (‘Umatilla Russet’ and ‘Alturas’) maturing cultivars were used. Yukon Gold is a yellow table stock cultivar, Désirée is a red table stock cultivar, and all of the other cultivars are russeted types used primarily in processing. In all trials, Pi/cultivar combinations were replicated at least seven times. All trials were arranged in a completely randomized block design in a single (Russet Burbank and Désirée trials) or multiple (cultivar and Ranger Russet trials) rows. Additional rows of potatoes were planted on either side of experimental rows to serve as an environmental buffer.

Prior to establishing treatments, the entire area was treated with the pre-emergence herbicide S-ethyl dipropylthiocarbamate (Eptam, Gowan Company, Yuma, AZ, USA) at a rate of 4.1 liter/ha and 1,344 kg/ha triple 16 (16-16-16 N-P-K) prior to shallow cultivation. To plant and inoculate potatoes, a 30-cm diam. × 30-cm deep hole, with the top of the hole at the base of the furrow, was excavated with a 30-cm diam. gasoline-powered auger. In-row spacing between plants was 76 cm with rows spaced 86 cm apart. To establish treatments, uninfested field soil was thoroughly mixed with pre-measured inoculum soil to produce the desired level of eggs/g soil for that Pi. Globodera ellingtonae inoculum soil with average egg densities of 219, 117, and 320 eggs/g soil and 259, 114, and 245 eggs/cyst were used in 2013, 2014, and 2015, respectively. A total volume of ~9,000 cm³ of mixed soil (with or without G. ellingtonae inoculum) was returned to augured holes. Cut tuber pieces, ~56 g, treated with mancozeb (Maxim MZ; Syngenta Crop Protection, Greensboro, NC) were then placed on top of the soil in the furrow. The area surrounding the experimental area was planted to Russet Burbank for a total of 16 rows. After hand planting, the tubers were treated with imidacloprid (Alias 4F; Adama, Aventura, FL) at 0.44 liter/ha and azoxystrobin (Quadris; Syngenta Crop Protection) at 0.9 liter/ha applied in a 15-cm band over the seed, and then the furrows were closed with additional soil using a tractor-mounted disk hiller. Potatoes were planted May 20, 2013, May 9, 2014, and May 14, 2015. The trial area was sprinkler irrigated and managed with cultural practices common in central Oregon. To control early blight (Alternaria solani), plants were treated with boscalid (Endura; BASF, Florham Park, NJ) at 0.33 liter/ha and azoxystrobin at 0.88 liter/ha based upon a growing degree day model used in the region (Gent and Schwartz, 2003). Weeds were additionally managed post-plant with pendimethalin (Prowl; BASF) at 2.34 liter/ha. Potatoes were monitored weekly for vigor including stunting and chlorosis.

Potatoes were harvested approximately 18-weeks-after planting. Aboveground biomass was collected 2 wk prior to harvest by clipping stems at surface level. Dry weights were determined after drying in a 50°C oven for at least a week. At harvest, tubers from each plant were hand-dug and individually weighed. At the same time tubers were collected, an approximately 500 g soil sample was collected near the root zone of each plant and placed in plastic bags. Soil samples were air dried on trays in a greenhouse for at least 1 wk and cysts extracted and enumerated as described above.

A one-way analysis of variance (ANOVA) was used to test for differences in mean yields between, inoculum levels in Trials 1, 2, 3, and 5 followed by a Tukey’s HSD test. To test for a relationship between Pi and yield, linear regression with log₁₀ (Pi + 1) was performed. Linear regressions were performed separately for Trials 1, 2, 4, and 5. A Welch two sample t-test was used to compare differences in mean yields between treatments for Trial 4. Host status and reproductive ability were measured by RF and Pf with differences among Pi densities tested by ANOVA and Tukey’s HSD. Data were log₁₀ ( x + 1) when necessary to meet the assumptions of ANOVA. All statistics were performed using JMP 13.

Results

Microplot experiments

Simple linear regression revealed that there was a positive correlation between logPi and tuber yield (Fig. 1); this correlation was significant (P ≤ 0.002). There was also a positive correlation between logPi and aboveground biomass (P ≤ 0.011, R ² = 0.08) (data not shown). The number of tubers produced per plant and the average individual tuber weight were not affected by increasing Pi (P ≥ 0.25) (data not shown). Number of tubers produced per plant ranged from 5 to 9 and average individual tuber weight ranged from 31.3 to 48.7 g.

Figure 1:

No cysts were recovered from the non-inoculated control in either trial (data not shown). Pf increased with increasing Pi (Table 2). While there was a significant difference in the number of G. ellingtonae eggs/cyst, for most Pi a similar number of eggs/cyst were produced (Table 2). RF values were significantly lower with higher Pi compared to lower Pi (Table 2).

Table 2.

Reproduction of Globodera ellingtonae inoculated at varying initial population densities (Pi) on potato (Solanum tuberosum) cv. Russet Burbank in microplots.

Field experiments

Tuber yields of Russet Burbank averaged 1,580 (±129) and 2,505 (±108) g/plant across all Pis in Trial 1 and Trial 3, respectively. The average tuber yield of Désirée in Trial 2 was 2,087 (±112) g/plant and of Ranger Russet in Trial 5 was 2,347 (±94) g/plant. Average per plant tuber yields for Russet Burbank were significantly different between Trial 1 and Trial 3, and between Désirée (Trial 2) and Russet Burbank (Trial 5) (P ≤ 0.01). Simple linear regression revealed that there was a negative correlation between logPi and tuber yield for Trials 1, 2, 3, and 5 (Fig. 2); however, the negative correlation was only significant for Russet Burbank in Trial 1 (P ≤ 0.015) and Ranger Russet in Trial 5 (P ≤ 0.005). In all of the experiments, a Pi of 80 eggs/g soil was included (Table 3). For the cultivar trial (Trial 4), no significant differences were observed between yields of inoculated (Pi = 80 eggs/g soil) and non-inoculated plants for any of the potato cultivars (P > 0.05) (Table 3). The only difference between yield of inoculated (Pi = 80 eggs/g soil) and non-inoculated plants was observed for Russet Burbank in Trial 1 and Ranger Russet in Trial 5 (P ≤ 0.05); there was a 43 and 19% reduction in yield of potato when G. ellingtonae was present compared to when the nematode was not present, respectively (Table 3). Comparing the Pf from the 80 eggs/g soil treatment among the potato cultivars (Trial 4) revealed significantly higher Pf for Russet Burbank than for three cultivars: Alturas, Umatilla Russet, and Yukon Gold (P ≤ 0.05) (Table 3). Likewise, the same significant differences between Russet Burbank and those three cultivars were observed for RF values (P ≤ 0.001). As another indicator of host quality, we compared the mean final eggs/cyst between all cultivars. All cultivars had significantly fewer eggs/cyst than Russet Burbank inoculated with 80 eggs/g soil (Table 3).

Discussion

This is the first experimental evidence regarding potential yield loss of potato caused by G. ellingtonae. High yield losses of potato caused by other species of PCN have been reported for decades, with as much as 80% yield loss at high nematode densities (Seinhorst, 1965; Brown, 1969; Greco and Moreno, 1992). Several studies report losses ranging from 25 to > 50% at Pi of 40 eggs/g soil and from 35 to 75% at Pi of 80 eggs/g soil, nematode densities similar to those tested in this study (Seinhorst, 1982; Elston et al., 1991; Greco and Moreno, 1992; Van den Berg et al., 2006). Similar yield loss to those reported for PCN might be expected for G. ellingtonae based on its phylogenetic proximity to PCN and high reproductive values on potato (Zasada et al., 2013); however, this level of loss due to G. ellingtonae was not consistently observed in two microplot and five field experiments conducted in Oregon.

Experiments with PCN have led to several proposed models to describe the relationship between Pi and potato yield (Seinhorst, 1965, 1982; Oostenbrink, 1966; Brown, 1969; Elston et al., 1991). These yield loss models vary greatly in their complexity, the more sophisticated of which include additional parameters such as minimum yield and damage thresholds. Separate experiments are required to determine those parameters for a new species at a new location, the experimental design for which would require far more inoculum than we were able to generate in the few years since the discovery of G. ellingtonae and the start of our research program with this nematode. Working within the constraints of our inoculum supply, we used mid-range initial nematode densities, shown both to be linearly correlated with yield and to have large negative effects on yield by PCN (Mulder et al., 1997), to determine if similar densities of G. ellingtonae would have a detectable effect.

As stated in the introduction, initially G. ellingtonae was found at very low densities (1-7 cysts/2.3 kg soil) across the farm at Powell Butte, OR, with one hotspot being identified with 24 cysts/2.3 kg soil. Therefore, two experimental approaches were undertaken to evaluate the pathogenic effects of G. ellingtonae. In the field experiments, we chose to utilize an artificial inoculation strategy. Similar to other studies, we inoculated the root zone with plant-parasitic nematodes and assessed plant growth in relationship to Pi (Thomas and Clark, 1983; Lamondia, 1995) We recognize the inoculation approach used could result in an underestimation of effect compared to a study in which the Pi was distributed over a larger volume of soil. Given that most root proliferation initially occurs near the plant (Kotcon et al., 1984; Opena and Porter, 1999) and the majority of G. ellingtonae eggs exposed to potato root diffusate hatch within seven days of exposure (Zasada et al., 2013), we expect our localized application of inoculum produced results similar to that observed with more distributed inoculum. In a study on the root distribution of potato in an irrigated field setting, the majority of the roots were found in the upper 20 cm of soil, with 3.9 cm roots/cm³ soil, compared to deeper in the soil profile down to 40 cm, with 0.9 cm roots/cm³ soil (Lesczynski and Tanner, 1976). In that study, horizontal distribution of potato roots at a depth of 10 to 20 cm increased with distance from the plant up to 25 cm from the plant. To further ensure we were encompassing potential pathogenic effects of G. ellingtonae on potato, we also conducted the microplot trials where the entire soil profile in which potato roots grew contained nematodes.

Field trials 1 and 5 were conducted with increasing Pi levels on Russet Burbank and Ranger Russet, respectively. Both trials showed a significant negative correlation between logPi and tuber yield. However, the linear models explained a notably small amount of the tuber weight variability (R ² = 0.23 for Russet Burbank and R ² = 0.20 for Ranger Russet). Based on the linear regression model of tuber yield on logPi for Trial 1, 30.5 to 40.9% yield loss was predicted at a Pi of 40 and 80 eggs/g soil, respectively, and for Trial 5, 16.5 and 19.7% yield loss was predicted at a Pi of 40 and 80 eggs/g soil. In the three other field trials and in the microplot trials, G. ellingtonae did not negatively affect tuber yield or plant growth. Additionally, no plant health issues (chlorosis, stunting, delayed germination) were observed on the potato plants in either the field or microplot experiments.

We already have reported that host resistance to G. ellingtonae varies with potato cultivar, generally in a pattern similar to G. rostochiensis (Zasada et al., 2013; Whitworth et al., 2018). However, host tolerance does not always directly correspond with host resistance (Trudgill, 1991). After observing a significant negative correlation between Pi and yield for Russet Burbank but no significant effect on Désirée in 2013, a trial was added in 2014 with different cultivars of potatoes to investigate whether maturity class might interact with effects on yield. Although some have found a correlation between maturity class and tolerance, others have not (Trudgill and Cotes, 1983; Arntzen and Wouters, 1994). None of the cultivars tested had significantly reduced yields compared to non-inoculated plants when inoculated with Pi of 80 G. ellingtonae eggs/g soil. Of the cultivars tested in the field, the host status of Désirée, Modoc, Umatilla, Russet Burbank, and Yukon Gold for G. ellingtonae had previously been evaluated in the greenhouse (Zasada et al., 2013). Consistent with greenhouse results, Yukon Gold had a lower, although not significantly so, RF value in the field than Russet Burbank, Umatilla, and Modoc. Globodera ellingtonae had higher Pf on Russet Burbank than Alturis, Umatilla Russet and Yukon Gold and higher eggs/cyst on Russet Burbank than all other six cultivars, consistent with previous observations (Zasada et al., 2013). This is notable as Russet Burbank is the most widely planted cultivar in the USA (National Agricultural Statistics Service, 2018).

Final egg densities in the field trials had maximum means ranging from 273 to 330 when inoculated with Pi of 40 eggs/g soil. In the microplot trials, Pf reached maximum means of 357 (Pi = 52) and 328 (Pi = 169) eggs/g soil. Combined, these results indicate successful invasion and reproduction of G. ellingtonae in all trials. Studies of PCN reproduction indicate the maximum Pf, often in a range of 300 to 400 eggs/g soil, usually occurs at intermediate Pi (~32–50 eggs/g soil) and that Pf values either level out or even decline with increasing Pi (Phillips et al., 1991; Greco and Moreno, 1992; Ehwaeti et al., 2000). The Pi of G. ellingtonae at which the maximum Pf occurred differed between years and between trials conducted within a year. As seen with PCN (Trudgill et al., 2014), RF values generally decreased with increasing Pi in our trials; the Russet Burbank 2014 trial was the only exception.

Field studies of PCN using the same potato cultivar for multiple years have documented that the rate of yield reduction varies with year, with effects of year sometimes eclipsing effects of Pi (Mulder et al., 1997). Our results are consistent with that pattern, with the difference in Russet Burbank yield between years 2013 and 2014 being much greater than the effect of G. ellingtonae in any given year. Although some studies have determined that relative yield loss caused by PCN increases with increasing total yield (Mulder et al., 1997; Phillips et al., 1998), we did not find such a correlation given that the only significant reduction in Russet Burbank yield was seen in the trial with a lower mean yield.

The pathogenic effect of G. ellingtonae on yield of potato was inconsistent across years and experimental venues. Five field trials were conducted over a three-year period and of these there was a significant negative relationship between G. ellingtonae Pi and tuber yield in the Russet Burbank Trial 1 and the Ranger Russet Trial 5. In two microplot experiments, there were no negative effects of varying G. ellingtonae Pi on tuber yield. Given that a field study comparing all three Globodera spp. at a single site is not feasible, using these data to compare pathogenic effects between species is not possible. Great care should be taken in extrapolating the results to probable effects in other environments, as effects of PCN on given potato cultivars vary when grown at different sites (Dale et al., 1988; Phillips et al., 1998). To date, G. ellingtonae has been reported from morphological descriptions only in Oregon and Idaho (USA) and Argentina, with presence in Chile inferred from molecular data (Skantar et al., 2011; Handoo et al., 2012; Lax et al., 2014). It will be of interest to determine the pathogenic effects of G. ellingtonae to potato in its presumably native environments of South America.